Credit Risk Meets Large Language Models: Building a Risk Indicator from Loan Descriptions in P2P Lending

Paper: arXiv 2401.16458

Authors: Mario Sanz-Guerrero, Javier Arroyo

Abstract

Peer-to-peer (P2P) lending connects borrowers and lenders through online platforms but suffers from significant information asymmetry, as lenders often lack sufficient data to assess borrowers’ creditworthiness. This paper addresses this challenge by leveraging BERT, a Large Language Model (LLM) known for its ability to capture contextual nuances in text, to generate a risk score based on borrowers’ loan descriptions using a dataset from the Lending Club platform. We fine-tune BERT to distinguish between defaulted and non-defaulted loans using the loan descriptions provided by the borrowers. The resulting BERT-generated risk score is then integrated as an additional feature into an XGBoost classifier used at the loan granting stage, where decision-makers have limited information available to guide their decisions. This integration enhances predictive performance, with improvements in balanced accuracy and AUC, highlighting the value of textual features in complementing traditional inputs. Moreover, we find that the incorporation of the BERT score alters how classification models utilize traditional input variables, with these changes varying by loan purpose. These findings suggest that BERT discerns meaningful patterns in loan descriptions, encompassing borrower-specific features, specific purposes, and linguistic characteristics. However, the inherent opacity of LLMs and their potential biases underscore the need for transparent frameworks to ensure regulatory compliance and foster trust. Overall, this study demonstrates how LLM-derived insights interact with traditional features in credit risk modeling, opening new avenues to enhance the explainability and fairness of these models.

Complexity vs Empirical Score

• Math Complexity: 3.0/10
• Empirical Rigor: 6.5/10
• Quadrant: Street Traders — practical and empirical, lighter on theory

Why this score: The paper provides a practical, empirical application of fine-tuning BERT to create text-based risk scores for P2P loan default prediction, subsequently feeding them into an XGBoost model. While mathematically straightforward (relying on standard ML architectures rather than novel mathematical derivations), it demonstrates solid empirical utility and performance gains on standard datasets.

Research Flowchart

  flowchart TD
    A["Research Goal: Mitigate P2P Lending Information Asymmetry using LLMs"] --> B["Data Input: Lending Club Loan Descriptions & Traditional Data"]
    B --> C["Methodology Part 1: Fine-Tune BERT on Loan Descriptions"]
    C --> D["Output 1: Generate Text-Based BERT Risk Score"]
    D --> E["Methodology Part 2: Train XGBoost Model with Traditional Features + BERT Risk Score"]
    E --> F["Key Outcomes: Improved Predictive Accuracy (AUC & Balanced Accuracy)"]
    F --> G["Key Insights: Uncovered Purpose-Specific Patterns & Feature Interactions"]

Cited References (63)

1. Attention is not Explanation (2019) — link
2. Post-hoc Interpretability for Neural NLP : A Survey (2022) — link
3. Is Attention Interpretable? (2019) — link
4. New explainability method for BERT -based model in fake news detection (2021)
5. Anchors: high-precision model-agnostic explanations (2018)
6. " Why Should I Trust You? “: E xplaining the Predictions of Any Classifier (2016) — link
7. Cost-sensitive boosted tree for loan evaluation in peer-to-peer lending (2017) — link
8. CatBoost: unbiased boosting with categorical features (2018)

0 of 63 cited works are in the corpus; resolved via arXiv source bundles, Semantic Scholar and OpenAlex. Corpus-wide: 2518 linked papers.